A retention curve looks simple — it's just a line on a graph. But the shape of that line contains more information about why your Shorts succeed or fail than any other metric in YouTube Studio.
This guide teaches you how to read a retention curve, identify what each shape means, and translate the data into concrete edits.
What a Retention Curve Is and How to Find It
A retention curve is a second-by-second graph of viewer retention. The Y-axis shows the percentage of your viewers still watching; the X-axis shows time in seconds. A 100% reading at the start means all viewers were watching. A 45% reading at 15 seconds means 55% had already left.
In YouTube Studio, find it under: your video → Analytics → Reach → Audience Retention. Select the "Absolute audience retention" option if given a choice between absolute and relative.
ClipHorizon automatically pulls and displays this curve for every Short connected to your channel, overlaid with timestamps for each significant drop and AI-generated explanations for each drop point.
The Six Retention Curve Shapes (And What Each Means)
1. The Ski Slope
Shape: Steady, continuous decline from the very first second to the end.
What it means: Viewers are leaving at a consistent rate throughout your entire video. There's no single catastrophic moment — just a slow bleed. This pattern usually means your content lacks urgency or visual variety. There's nothing compelling enough to make viewers actively choose to keep watching.
The fix: Add a pattern interrupt every 5–8 seconds. A new on-screen text overlay, a cut, a change in music, or a visual element that resets the viewer's attention. The goal is to create micro-reasons to stay at regular intervals rather than relying on narrative momentum alone.
2. The Cliff
Shape: Flat or gently sloping, then a sudden steep vertical drop at one specific moment.
What it means: Something specific happened at that timestamp that caused a large number of viewers to leave simultaneously. This is often a topic change, a visual disruption, a drop in energy, or the resolution of a setup (once viewers got the answer, they left). The cliff is the most useful retention shape to diagnose because it points to a specific, fixable moment.
The fix: Go to the exact timestamp of the cliff. Watch what was happening at that second. Common causes: a long pause, cutting to B-roll that feels disconnected, transitioning from the hook's promise to a dry explanation, or the video delivering its main payoff with minutes still remaining. Fix the specific moment, not the overall structure.
3. The Early Exit
Shape: A dramatic drop in the first 2–3 seconds, then the curve stabilizes.
What it means: Your hook is the problem. The viewers who make it past the first few seconds are engaged — but a large portion swipes before they've committed. This is the most common retention failure for YouTube Shorts. It corresponds directly to a low hook score.
The fix: Rewrite the opening entirely. Your first frame and first audio line are doing all the work. Common fixes: remove any introduction or greeting, start mid-sentence or mid-action, add a compelling text overlay in the first frame, or lead with your video's most surprising or counterintuitive claim.
4. The Plateau
Shape: A sharp drop in the first few seconds (typical hook phase), then an extended flat section through most of the video.
What it means: After the initial hook drop, the remaining audience is highly engaged and staying consistently. This is a decent outcome — you lost some viewers early, but those who stayed were committed. The hook needs work, but the content itself is strong.
The fix: Focus almost entirely on the hook. If the curve flattens after the 10% mark, your content is working well. Improving the opening 5–10 seconds will increase the number of viewers who reach the engaged plateau — and directly increase both your hook score and your overall view count.
5. The Terminal Rise
Shape: The curve dips and then rises at the very end, sometimes above the preceding seconds.
What it means: Viewers are rewatching. When the terminal value rises above the preceding seconds, it means more viewers are at that timestamp than were a few seconds earlier — which can only happen through replays. Rewatches are among the strongest signals the YouTube algorithm uses to identify high-quality content.
The fix: This is a positive pattern — reinforce it. Study what's in the last few seconds of your highest-retention Shorts. Usually it's either an unexpected payoff, a callback to the opening hook, or a cliffhanger that makes the viewer want to restart. Build this structure deliberately into future Shorts.
6. The Perfect Curve
Shape: A minimal initial drop (hook retention of 85%+), a relatively flat middle, and a terminal rise at the end.
What it means: Your Short is working at every stage. Viewers are committing to the opening, staying through the middle, and rewatching the end. This is the curve shape that consistently earns wide distribution and appears in the Suggested feed.
The fix: Document exactly what you did in this video — the hook structure, the pacing, the visual variety, the ending type. This is your template. Replicate the pattern, not the content.
Reading the Numbers: What Each Percentage Means
It's not just the shape but the specific values that matter.
Opening value (0:00–0:02): Should be 95–100%. If it's below 90% in the first two seconds, viewers are swiping before your hook even has a chance to land.
Hook phase end value (first 30% of video): This is your hook score as a raw percentage. Above 80% is excellent; 65–79% is good; below 55% is problematic.
Midpoint value (50% of video length): A healthy Short retains 55–70% of viewers at the midpoint. Below 45% at the midpoint means you're losing viewers faster than the algorithm expects.
Terminal value (final 5% of video): A terminal value higher than the few seconds preceding it indicates rewatches. Even a 2–3% rise is meaningful signal.
Overall average (the number YouTube Studio reports): This averages all the above into a single number. Aim for 60%+ as a general target, with 80%+ as the goal for broad distribution.
The Five Timestamps to Check First
When you pull a new Short's retention curve, check these five moments before analyzing anything else:
Seconds 0–3. Is there an early exit? If retention drops below 75% in the first three seconds, your hook failed and that's your highest-priority fix.
The 10% mark. Where is retention after the initial hook phase? If it's below 65%, the hook structure is weak even if there wasn't a dramatic spike at the very start.
The 50% mark. Is retention still above 50% halfway through? If not, there's a pacing or content structure problem in the middle section.
Any cliff. Look for any point where the line drops steeply in a short timeframe. These are your specific fix points.
The final 5%. Is there a terminal rise? If not, your ending isn't earning rewatches. Try a different ending structure — a callback, a surprise reveal, or a question that makes the viewer want to restart.
How ClipHorizon Uses Retention Curves
ClipHorizon pulls the second-by-second retention data from YouTube Analytics API and overlays it with AI analysis. For each significant drop point, it generates:
- A timestamped explanation — what was happening at that exact second and why viewers likely left
- An edit suggestion — a specific, actionable change to prevent that drop in future Shorts
- A hook rewrite — an alternative opening line designed to improve hook retention based on your actual video content
- A hook score — the 0–100 metric derived from the hook phase retention data
This turns the retention curve from a passive data display into an active editing guide.
Frequently Asked Questions
What is a YouTube Shorts retention curve?
A YouTube Shorts retention curve is a second-by-second graph that shows what percentage of viewers are still watching at each point in your video. The shape of the curve tells you where you lost viewers, where you kept them, and where viewers rewatched — all of which affect how the YouTube algorithm distributes your Short.
What does a drop on a retention curve mean?
A drop means that a measurable percentage of viewers stopped watching at that exact second. The steeper the drop, the more viewers you lost in that moment. Drops in the first 3 seconds indicate a weak hook. Drops at the midpoint often indicate pacing problems. Drops at the end usually indicate a weak payoff.
What is a good retention curve shape for YouTube Shorts?
The ideal shape is relatively flat from beginning to end, with a slight rise at the end (indicating rewatches). The worst shapes are the Ski Slope, the Cliff, and the Early Exit. The best-performing Shorts have minimal hook drop, a stable middle, and a terminal rise from loop views.
Where can I see my YouTube Shorts retention curve?
In YouTube Studio under Analytics > Reach > Audience Retention for each video. ClipHorizon automatically displays the retention curve for every analyzed Short alongside timestamped drop-off explanations and AI-generated edit suggestions for each significant drop point.